[Tig] color nomenclature tool

Richard Kirk richard at filmlight.ltd.uk
Thu Nov 5 10:52:17 GMT 2009



Rob Lingelbach <rob at colorist.org> writes...
> To: "tig at colorist.org Group" <tig at colorist.org>
> Subject: [Tig] color nomenclature tool
> Message-ID: <576F119E-69A5-4D25-8575-99A4601BF37A at colorist.org>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed; delsp=yes
>
> I wanted to incorporate on the TIG wiki one of the better tools for
> color nomenclature,
> which includes a color picker with RGB or HSB selection and readout
> and web hex color, but most importantly, one with a very large color
> dictionary.  One of the best ones is Daniel Flück's
> work which he graciously gave me permission to use in the TIG wiki.
>
> http://tig.colorist.org/wiki3/index.php/Color_Picker
>
> I find the color naming conventions as interesting as the actual
> variations; who determined that a color of R=0 G=88 B=96 should be
> called Mosque?  or that G=76 B=79 should be Sherpa Blue?  Which I
> wouldn't mind in a Ferrari Dino.

No Cambridge Blue? Bah.

> There are no explicit or implied warranties regarding the accuracy of
> the colors, considering all the possible variables between the
> original coding and the eventual display device and eye of beholder.

True. We don't know the display or the tone curve or the viewing
conditions. However, things are not necessarily that bad. If the display
is self-luminous then the user can look at it in reduced lighting if they
want to be accurate. Dark is the same everywhere, so that ought to work.
We are then looking at the color on a self-luminous screen with other
stuff that gives us the sense of display white. That is not a bad
criterion for viewing conditions.

We do not know the monitor white, but most monitor whites are about D65
these days. There is no simple test for D65.

The monitor gamma is not specified. However, there are a set of gamma
tools at the bottom of...
http://en.wikipedia.org/wiki/Gamma_correction

I have an image which can be used to estimate monitor gamma at highlights,
mid-tones and shadows (I will post it if I can find it). If we agreed a
gamma, then this would determine the red, green, and blue brightness
ratios for the higher values. Down in the low values, display gammas are
not constant (they can increase sharply or flatten off to zero) so very
dim colors can probably not be trusted.

We do not know the monitor RGB primaries. However, if R=G=B looks neutral
and the monitor can display other images correctly, then they are probably
close to the standard ITU or EBU phosphor sets. If we increase the monitor
brightness, our sense of color increases (the Hunt effect). If we increase
the saturation, our sense of brightness can increase (the Stevens effect).
In practical terms, we are probably ranging our sense of color to fit what
we see, and we will not spot the difference between a video display and a
video+10% display unless they are put next to each other. I vividly
remember seeing the same video signal going to a Sony BVM and a Barco
DP90P, and the red on the Sony looking red until the Barco turned on, when
it promptly turned orange.

In this case, the worded description of a color is probably a fair match
to our internal sense of that color if we do not have any color stimuli
coming from outside the display. Of course, that is an easy thing to say
as we have no way of measuring our internal sense of color. I suspect the
tool would work well enough.

But, is it useful?

Cheers.
Richard Kirk







More information about the Tig mailing list